AW S
Hajime Sano, Marketing & Data Technologist
Data Analytics & CRM Center, B to C Unit, Nikkei Inc.
1.
2.
3. TIPS
•
• AWS 

DynamoDB DAX
•
• www.linkedin.com/in/hsano
1.
2.
3. ≠
1.
2.
3. TIPS
+
frequency √ volume
1 2 3 4 5 6 7 8 9 10 11 12 13 14
= LOG2(F√V)
• 10
•
•
•
• FT …
• GROUP BY 50
•
• Latency
•
•
2014 2015 2016 2017 2018
•
•
•
•
AT L A S …
•
•
•
• 10 1
•
4
R D B
H A D O O P
B I
D A S H B O A R D
R D B
H A D O O P
B I
D A S H B O A R D
1.
2.
3. TIPS
EndpointTracking Enrichment Consumers S3 AzkabanSQS Kinesis S3 Redshift
Consumers ESS3
ParserAdobe Analytics
Dynamo DB Dynamo DB
S3
Kinesis→ES
DataFeed
Kinesis→S3 S3→RS
E B / E C 2
Rundeck
Kinesis S3 Redshift
Athena
Spectrum
ES Kibana
Analytics
Firehose
Quick Sight
(R)
Jupyter
(py)
OSS BI/DS
B I
AW S
SQS Kinesis S3 Redshift
S Q S
•
-
-
-
K I N E S I S
• Kinesis Stream
-
- 7
-
• Firehose Analytics
S 3
•
• Redshift
• S3
- Redshift Athena Spectrum
-
R E D S H I F T
•
• postgres SQL
- BI
•
-
E L A S T I C S E A R C H E C 2
• Elasticsearch Service EC2
- c4.8xlarge x 3
- r4.xlarge x 1
•
- X-Pack Graph ML
- ES OS JVM
• Kinesis
- 200ms
- Consumer
• Elasticsearch
-
- INDEX 1
• Redshift
- 15 20
10 20
10 1
…
AT L A S
A
G
1.
2.
3. TIPS
•
• Lightning Talk
• SQL Data Dojo
•
•
• / IF
R Studio Server Shiny Anaconda(JupyterHub)
Chartio
Re:dash Kibana
DOMO
DataSquad Maia
KPI Screens
• 2
1.
2.
•
- Slack
-
J O I N - L E S S
• OUTER JOIN
-
- JOIN
• 1
- 1
- SELECT
J O I N - L E S S
GET URL URL URL
•
- time sliced table UNION ALL
- timestamptz
- WHERE
•
- 3 30 …
-
last 7 days 7
•
- Redash
…
-
•
- Redash
- 

KILL
T I P S
1.
2.
3.
4.
T H A N K Y O U !
Data Technologist Data Scientist
Q U E S T I O N S ?

リアルタイムアクセスログ分析基盤をAWSに構築した話 (JAWS UG BigData Branch)